Available Models

ProxAI provides functions to list and filter available models based on your configured API keys.

Examples

Simple Usage

import proxai as px
 
# List all available models
models = px.models.list_models()
 
for model in models:
  print(model)

(anthropic, claude-sonnet)
(openai, gpt-4)
(openai, gpt-4o)
(google, gemini-2.0-flash)
...

Filter by Features

import proxai as px
 
# List models that support web search and pydantic response format
models = px.models.list_models(
    features=['web_search', 'response_format::pydantic'])
 
for model in models:
  print(model)

(openai, gpt-4o)
(openai, gpt-4o-mini)
...

Filter by Model Size

import proxai as px
 
# List only large models
models = px.models.list_models(model_size='large')
 
for model in models:
  print(model)

(anthropic, claude-sonnet)
(openai, gpt-4o)
(google, gemini-2.0-flash)
...

Functions

There are two categories of functions for checking models:

List all configured models (based on API keys in environment):

• px.models.list_models()
• px.models.list_providers()
• px.models.list_provider_models()
• px.models.get_model()

List verified working models (tested with actual API calls):

• px.models.list_working_models()
• px.models.list_working_providers()
• px.models.list_working_provider_models()
• px.models.get_working_model()

px.models.list_models()

It is possible to list all available models as follows:

provider_models = px.models.list_models()
 
for provider_model in provider_models:
  print(provider_model)

(claude, haiku)
(claude, opus)
(claude, sonnet)
(gemini, gemini-2.0-flash)
(gemini, gemini-2.0-flash-lite)
(gemini, gemini-2.5-pro-preview-03-25)
(openai, chatgpt-4o-latest)
(openai, gpt-4.1)
...

• Returns a list of ProviderModelType objects.

Parameters

px.models.list_providers()

Lists all providers that have API keys configured in environment variables.

providers = px.models.list_providers()
 
for provider in providers:
  print(provider)

anthropic
google
openai
...

• Returns a sorted list of provider names (List[str]).

Parameters

px.models.list_provider_models()

Lists all models available from a specific provider.

provider_models = px.models.list_provider_models(provider='openai')
 
for provider_model in provider_models:
  print(provider_model)

(openai, gpt-4)
(openai, gpt-4-turbo)
(openai, gpt-3.5-turbo)
...

• Returns a list of ProviderModelType objects.

Parameters

px.models.get_model()

Gets a specific model by provider and model name. This function returns the model if it exists and the provider’s API key is configured, but does not verify that the model is actually working.

model = px.models.get_model(provider='openai', model='gpt-4')
 
print(model)

(openai, gpt-4)

• Returns a ProviderModelType object.
• Raises ValueError if the provider’s API key is not found or the model doesn’t exist.

Parameters

px.models.list_working_models()

Lists models that have been verified to work by sending actual test requests. Results are cached to avoid repeated testing.

working_models = px.models.list_working_models()
 
for model in working_models:
  print(model)

(anthropic, claude-3-opus)
(openai, gpt-4)
(openai, gpt-3.5-turbo)
...

• Returns a list of ProviderModelType objects, or a ModelStatus object if return_all=True.

Parameters

px.models.list_working_providers()

Lists providers that have at least one working model. Tests models and returns providers that have successfully responded to at least one test request.

providers = px.models.list_working_providers()
 
print(providers)

['anthropic', 'openai']

• Returns a sorted list of provider names (List[str]).

Parameters

px.models.list_working_provider_models()

Lists working models from a specific provider by testing each model.

openai_working = px.models.list_working_provider_models('openai')
 
for model in openai_working:
  print(model)

(openai, gpt-4)
(openai, gpt-3.5-turbo)

• Returns a list of ProviderModelType objects, or a ModelStatus object if return_all=True.
• Raises ValueError if the provider’s API key is not found.

Parameters

px.models.get_working_model()

Verifies and returns a specific model if it’s working. Tests the model and returns it only if it successfully responds.

model = px.models.get_working_model('openai', 'gpt-4')
 
print(model)

(openai, gpt-4)

• Returns a ProviderModelType object if the model is working.
• Raises ValueError if the provider’s API key is not found, the model doesn’t exist, or the model failed the health check.

Parameters

Details of how models are checked

The available models are determined as follows:

• It checks environment variables for provider API keys (see Provider Integrations)
• Then, it looks if the model was previously cached
• Finally, it makes a simple request for each un-cached model to check if these models are available